Fishing normally begins in this area of 

 the coast during the first week in June and ter- 

 minates during the first week in October. 

 Irregularities in the movements and behavior 

 of the fish, weather, and other factors, may 

 extend either end of the season (fig. 5). To 

 eliminate those scattered landings which pre- 

 ceded and followed the period when the bulk of 

 the catch was made each year, the period June 

 1 through October 4 arbitrarily was chosen for 

 the analysis. For the period of this study, the 

 landings outside the season, thus defined, 

 amounted to no more than 9.3 percent of the 

 total catch in any year (table 1) . The season 

 June 1 - October 4 was divided into 7 -day periods 

 (boat -weeks) and comparisons made of each 7- 

 day period with the corresponding period in the 

 following season. There were 18 such time 

 units in the season. 



It will be recognized that great variability 

 exists in the size and efficiency of a group of 

 vessels over a period of time. It was desirable 

 therefore, to employ a catch -per -unit index 

 which tended to minimize these factors as much 

 as possible. One such measure would involve 

 the selection of those vessels which had fished 

 continuously over the 15-year period. Because 

 vessels were leaving and entering the fishery 

 throughout the period, however, it was impos- 

 sible to obtain a sufficient number to yield a 

 reliable measure. 



To establish criteria for the selection of 

 boats, several conditions had to be considered. 

 In certain years some vessels did not fish the 

 entire season due to breakdowns or other 

 reasons independent of the availability of fish. 

 To minimize possible bias, vessels which landed 

 fish in less than one -half the number of boat- 

 weeks in a season were omitted. Furthermore, 

 there is no sure way of ascertaining whether a 

 boat was fishing during periods when it made no 

 landings. It was found, however, that the bulk 

 of zero catches occurred at the beginning or 

 end of a given season, suggesting the vessels 

 were not fishing during the period in question. 

 Finally, the number of zero catches during any 

 given season being small, no adjustment was 

 made for these values. This resulted in the 

 eliminating of no more than 4.0 percent of the 

 data for any pair of seasons over the period 

 considered. 



Applying the time unit (boat -week) and 

 the criteria for selecting vessels in the fleet, 

 it is possible to make a series of comparisons 

 for each of the pairs of seasons for the period 

 1939-1953. 



The total tonnages delivered by all boats 

 in the selected fleet each season were summed. 

 The sum for the second of a given pair of years 

 was divided by the sum of the first year yielding 

 the ratio of the second year to the first. The 

 total tonnage was then divided by the total boat- 

 weeks to obtain the catch per boat -week for 

 each season. 



A series of indices of catch per boat- 

 week was obtained using 1946 as the base year. 

 By means of the ratios, other years were linked 

 to 1946 and the index values determined. The 

 calculated values are given in table 2 and shown 

 graphically in figure 6 . 



Table 2 , 



■-Catch per unit effort and fishing 

 effort, menhaden purse seine 

 fishery, 1939-1953. 



The calculated catch per unit effort 

 (fig. 6) shows a general upward trend from 1939 

 to 1947, then decreased until 1951. This was 

 followed by a rather marked upward trend in 

 1952 and 1953. The calculated catch per boat- 

 week in 1946 and 1947 (447 and 460 tons 

 respectively) was more than double that of 1939 



10 



